Synthesis Types
"Synth" covers many different ways of making sound. Some carve tone out of a bright wave, some build it up from sine waves, some bend one oscillator with another, and some chop recordings into tiny grains. Knowing the main types tells you which instrument to open for the sound in your head.
Subtractive Synthesis
Subtractive synthesis starts with a harmonically rich waveform (saw, square, noise) and removes frequencies with a filter. It is the method you learned in Synth Basics and Filters, and it is still the most common approach.
Classic examples are the Minimoog, the Roland Juno series and the Sequential Prophet-5. Most stock DAW synths (Ableton Analog, FL Studio 3x Osc, Logic Retro Synth) are subtractive.
- Strengths: easy to learn, warm, great for basses, leads, pads and brass.
- Limits: hard to make bells, metallic tones or realistic acoustic instruments.
Additive Synthesis
Additive synthesis works the other way round. It builds a sound by adding together many sine waves, called partials. You set the frequency and level of each one. This rests on the idea (from Fourier) that any periodic sound can be described as a sum of sine waves.
The Hammond organ is an early mechanical example: each drawbar adds a sine-like tone at a different harmonic. Software such as FL Studio's Harmor and Logic's Alchemy include additive engines, and many of them can resynthesize a sample by analyzing its partials.
Building a square wave with additive
Add enough odd harmonics at these levels and the result approaches a square wave. Change the levels over time and you get sounds no filter could make.
- Strengths: precise control of every harmonic, smooth morphing, organ and glassy tones.
- Limits: many parameters, so it is slow to program by hand.
FM Synthesis
FM (Frequency Modulation) synthesis uses one oscillator to change the frequency of another at audio rate. If an LFO on pitch gives vibrato, FM is vibrato so fast (hundreds or thousands of times a second) that you hear it as new harmonics instead.
- The carrier is the oscillator you hear.
- The modulator is the oscillator that bends the carrier's frequency.
- The ratio is the modulator's frequency compared with the carrier's. Whole-number ratios (1:1, 2:1, 3:1) give harmonic, musical tones. Non-whole ratios (such as 1:1.41 or 1:3.5) give inharmonic tones: bells, metal, clangs.
- The modulation index is how strongly the modulator bends the carrier. More index means more harmonics and a brighter sound. Putting an envelope on the index makes a note start bright and then mellow, just like a struck bell or piano.
In FM synths, each oscillator is called an operator, and the way operators are connected is called an algorithm.
John Chowning developed FM synthesis at Stanford University and the patent was licensed to Yamaha. The Yamaha DX7 (1983) brought it to a mass market. It has six operators and 32 algorithms. Its electric piano, bell and slap bass sounds were all over 1980s pop. Technically Yamaha's chips use phase modulation, a close relative that behaves almost the same. Modern FM synths include Ableton Operator, FL Studio Sytrus, Native Instruments FM8 and Arturia's DX7 V.
Now the electric piano. The ratio is 1:1 and the index is low, so the tone is warm with a bright "tine" at the start that fades out as the index decays.
- Strengths: bells, electric pianos, plucked basses, metallic and glassy tones, aggressive growls in modern bass music.
- Limits: small changes have big, sometimes unpredictable results. It takes practice.
Wavetable Synthesis
Wavetable synthesis uses a wavetable: a stack of single-cycle waveforms stored one after another. The wavetable position control picks which frame plays. Moving the position (often with an LFO or envelope) morphs the sound smoothly from one waveform to the next.
A wavetable
The output then usually goes through a filter and amp, just like subtractive synthesis. Wavetable is "subtractive with moving oscillators".
Wavetable began with Wolfgang Palm's PPG Wave synths in the early 1980s and continued with Waldorf instruments. Today the best-known software is Xfer Serum and Vital (which has a free version), plus Ableton Wavetable. Both Serum and Vital let you import your own audio as a wavetable. They are the standard tools for modern EDM leads and bass.
- Strengths: evolving pads, growling basses, modern EDM sounds, visual and fast to program.
- Limits: can sound digital or harsh if you are not careful with the filter and EQ.
Granular Synthesis
Granular synthesis chops a sample into tiny pieces called grains, usually 1–100 ms long, and plays many of them at once. Key controls:
- Position: where in the sample the grains are taken from.
- Grain size: the length of each grain. Short grains sound buzzy. Long grains keep more of the original.
- Density: how many grains play per second.
- Spray or jitter: random variation in position, pitch or pan for each grain.
Freeze the position on one spot and a single word of a vocal becomes an endless pad. Granular engines appear in Logic Alchemy, Arturia Pigments and Ableton's Max for Live Granulator devices. It is a staple of ambient music and cinematic sound design.
Physical Modelling
Physical modelling uses math to simulate how a real object vibrates: a string being plucked, air in a tube, a mallet hitting a bar. Instead of "cutoff" you get controls like string stiffness, pick position, body size and damping.
The simplest example is the Karplus-Strong algorithm: a burst of noise fed into a short delay line with feedback and a filter sounds surprisingly like a plucked string. Commercial examples include the Yamaha VL1 (1990s), Ableton Tension (strings) and Collision (mallets), and Modartt Pianoteq (pianos).
- Strengths: very expressive, responds naturally to velocity, small file size.
- Limits: can sound "almost real but not quite", and it is CPU-heavy.
Sample-Based Synthesis & ROMplers
Sample-based instruments play back recordings of real instruments and then shape them with filters, envelopes and LFOs. A ROMpler is a sample-based instrument with a fixed library of sounds (originally stored on ROM chips). The Korg M1 (1988) and the Roland JV series were hugely popular hardware ROMplers.
Today's software samplers include Native Instruments Kontakt, Spectrasonics Omnisphere (sample-based plus synthesis), and stock tools like Ableton Sampler, FL Studio DirectWave and Logic Sampler. See Sampling for creative sample use.
- Strengths: the most realistic acoustic instruments (pianos, strings, choirs, drums).
- Limits: large libraries, and less freedom to reshape the core sound.
Which Type for Which Sound?
| Sound you want | Best first choice | Why |
|---|---|---|
| Fat analog bass or lead | Subtractive | Saw waves plus a resonant filter are the classic recipe |
| Warm pad | Subtractive or wavetable | Detuned saws for warmth; wavetable for slow movement |
| Bells, chimes, glassy keys | FM | Non-whole ratios create inharmonic partials |
| 1980s electric piano | FM | The DX7 sound itself |
| Growling EDM bass | Wavetable or FM | Moving wavetable position or index gives aggressive motion |
| Organ | Additive | Drawbar-style harmonic control |
| Ambient texture from a vocal or field recording | Granular | Stretches and freezes real audio |
| Expressive plucked string or mallet | Physical modelling | Responds like the real object |
| Realistic piano, strings or orchestra | Sample-based | Recordings of the real thing |
Modern synths are often hybrids. Serum, Vital, Pigments and Alchemy mix wavetable, FM, sample and granular engines in one instrument. Don't worry about labels. Learn what each engine is good at, then combine them.
- Open your DAW's FM synth [Ableton: Operator; FL Studio: Sytrus or 3x Osc with FM; Logic: Retro Synth FM mode or ES2 FM] and load an init patch.
- Use two operators: B (modulator) feeding A (carrier). Set both to sine.
- Set B's coarse ratio to
1and slowly raise its level (the index). Listen to harmonics appear. - Change B's ratio to
2, then3, then a fine-tuned value like3.5. Note which settings sound musical and which sound like bells. - Give operator B a short envelope (attack 0 ms, decay 400 ms, sustain 0%). The note now starts bright and mellows: an FM pluck or e-piano.
- Now open a wavetable synth [Ableton: Wavetable; FL Studio: FLEX or Vital as a plugin; Logic: Alchemy or Vital] and assign a slow LFO (4 bars, synced) to wavetable position. Compare how the two engines create movement.
In FM synthesis, the oscillator you hear is the and the one that bends it is the .
The famous Yamaha FM synth from 1983 is the .
Granular synthesis splits a sample into tiny pieces called .
1. Which synthesis type removes harmonics from a rich waveform with a filter?
- Subtractive
- Additive
- Granular
- Physical modelling
Subtractive synthesis starts bright and subtracts frequencies with a filter.
2. In FM synthesis, what usually happens when you use a non-whole-number ratio like 1:3.5?
- The sound gets quieter
- The sound becomes a pure sine wave
- The tone becomes inharmonic, like a bell or metal
- The pitch drops an octave
Whole-number ratios give harmonic tones. Non-whole ratios create partials that don't line up with the harmonic series.
3. What does the wavetable position control do?
- Sets the filter type
- Moves the sound in the stereo field
- Changes the sample rate
- Chooses which waveform frame in the table is playing
A wavetable is a stack of waveforms. Moving the position morphs between them.
4. You want a realistic grand piano for a ballad. Which type is usually the best first choice?
- FM
- Wavetable
- Sample-based
- Granular
Sample-based instruments play recordings of real pianos, which is still the most realistic option. Physical models like Pianoteq are a strong alternative.
5. Which synthesis type builds sounds by adding sine waves together?
- Subtractive
- Additive
- Sample-based
- Physical modelling
Additive synthesis sums sine-wave partials, each with its own frequency and level.